The Deaths of Ideas in Congress

Abstract

Research on lawmaking generally examines why lawmakers enact certain policy ideas. However, most ideas in the United States Congress never become law. Eventually, members stop introducing them altogether. This process, where policies are proposed, not acted on, and no longer advocated for is, by far, the most common legislative outcome. The literature pays it little attention. Yet, how ideas die is important in understanding why some measures stop being realistic alternatives, the importance of policy entrepreneurship, and how policy windows affect the supply of ideas. This paper analyzes why congressional ideas die. I argue and find that the proposals designed to be enacted, especially by legislators with more issue expertise and agenda-setting powers, are less likely to persist across terms. I also show that ideas disappear more often when their sponsors leave Congress, but do not find a similar pattern for when policy windows close and lawmaking conditions worsen.

Keywords

Congress lawmaking agenda-setting text analysis

Introduction

Who or what is killing policy ideas on Capitol Hill? Although most attention is paid to measures that fail loudly, by filibuster, the veto, or inter-chamber disagreements, most quietly disappear. They are simply not reintroduced in the next congressional term. This process, in which an idea is proposed, not enacted into law, and then not proposed again is, by far, the most common life cycle for a congressional measure. Yet, we know very little about why legislators continue to reintroduce policy ideas or let them disappear from the legislative agenda.¹

In one sense, this lack of scholarly attention makes sense. Understanding how an idea is enacted, and comes to affect people’s lives, is important. By definition, measures that are not enacted, especially those no longer considered, are functionally irrelevant. Yet, understanding why ideas disappear is important for two reasons. First, an idea must first be introduced to be enacted. Although seemingly obvious, the legislative agenda’s scope defines what is possible. Once an idea leaves it, the measure is no longer a viable option.

Second, the reasons why ideas disappear speak to important questions about influence and power in Congress. Research on lawmaking and descriptive representation emphasizes the importance of policy entrepreneurs bringing new ideas to Congress (Kingdon 2003; Schiller 1995). Indeed, a significant amount of research, particularly on descriptive representation, contends that members’ backgrounds shape their legislative proposals (Griffin 2014). Yet, this connection between members being the only or main champions of an idea is largely assumed. Political scientists know very little about how much a measure is tied to a specific legislator and even less about its durability once their main advocate leaves office. Studying the process by which an idea leaves the agenda offers insight about how closely a proposal is tethered to an individual member and the importance of that initial entrepreneurship. Put differently, once introduced, does the measure gain a life of its own?

Similarly, theories of agenda-setting and legislative effectiveness emphasize that introduced ideas rely on either having supporters build momentum or waiting for the right moment (Baumgartner and Jones 2009; Kingdon 2003; Volden and Wiseman 2014). Although many case studies show that some proposals linger in Congress for years and eventually become law, most do not (Rawat and Morris 2016). Instead, they die. Put differently, Kingdon (2003) describes enacted policies as “ideas whose time has come.” For most proposals, the time never comes.

In this paper, I examine why ideas disappear from the congressional agenda. Disappearances are seemingly puzzling because a reintroduction has very low costs but produces position-taking benefits. Most of the costs to develop a measure are sunk. The actual mechanics of keeping it alive—minor revisions to already written legislative text or recycling press releases—are cheap. Even with this seemingly favorable cost-benefit calculation, most ideas are not reintroduced in the subsequent congressional term.

To address this puzzle, I consider the more complex trade-off members face when deciding to reintroduce a measure or not. In particular, I account for how the maintenance costs of keeping an idea alive and legislators’ informational disparities about its likelihood of enactment affects their decision. I draw on ideas from the agenda-setting literature that emphasize legislators must continually work to position their ideas as a workable alternative to a problem, even if that issue is not being immediately addressed (Baumgartner and Jones 2009; Kingdon 2003).

When deciding whether to drop an idea or not, I argue that a lawmaker’s decision is based on whether the proposal’s continued value exceeds what they can gain from introducing a new measure. By clarifying the varying costs and benefits that members face for various types of legislation, new or a retread, I contend that for some members it is easier to offer a different one. In particular, those whose ideas are referred to their committees, and especially committee leaders, face lower introduction costs and have better information about a current measure’s enactment prospects. This makes it easier to kill an idea and introduce something new once it becomes clear the measure is unlikely to pass. Thus, counter-intuitively, those who are best positioned to have their ideas put into law are the least likely to keep them alive across congressional terms.

I test this argument by examining which policy ideas, operationalized as legislative sections, are not reintroduced in subsequent congressional terms from 1993 through 2016. Using text reuse methods and a decision tree machine learning model to match sections that contain the same policy proposal, I track which measures are reintroduced or disappear from the agenda. I find consistent evidence that legislators are more likely to let measures die on issues they know more about and have jurisdiction over. I consider other explanations, such as ideas disappear when members leave Congress, policy windows close, or lawmaking conditions worsen. Although a sponsor departing Congress increases the probability an idea disappears, it is not determinative. I find no evidence for alternative explanations, such as closing policy windows, worsening lawmaking conditions, or that some members simply have larger legislative portfolios.

By examining this very common but rarely studied process, my findings provide new insights about lawmaking. Better resourced legislators, especially on the topics they have more control over, are more likely to drop proposals. Members are tied to their ideas, but sometimes a measure persists after its sponsor leaves office. More commonly it dies while they remain in Congress. Together, these results suggest that a proposal’s perceived viability is a poorly understood aspect of legislating. Research that highlights how measures build support over time select on this concept and gloss over why a member believed a proposal was worth keeping alive in the first place. As such, future studies should explore why such a small subset of ideas are kept alive over the long term.

Background

When learning about the US Congress, students are taught that most measures are not enacted. What happens to these tens of thousands of policy ideas that do not become law every congressional term? One possibility is that they are reintroduced and linger on the agenda, waiting for the right moment to be enacted. This process, although uncommon, has been extensively studied (for reviews, see Kuhlmann and van der Heijden 2018; Rawat and Morris 2016). The second option is that they stop being proposed and disappear from the agenda, a process that describes most measures and has received very little attention.

From a cost perspective, it is puzzling why a member would choose to let an idea die rather than reintroduce it. It is already written and when proposing it the first time, the legislator went to the trouble of finding cosponsors, seeking support among interest groups, and developing a communications strategy to claim credit. Those sunk costs have already been paid. In comparison, a reintroduction is essentially free. A member only needs to, at most, update some legislative language, walk the legislation to the floor, sign it, and resend their already written constituent communications.

Given this ease, most research treats an idea’s legislative failure as an intermittent step in eventually becoming law. In particular, studies on agenda-setting highlight how proposals linger as their sponsors begin building support for them while waiting for the right moment to push for their enactment (Baumgartner and Jones 2009; Kingdon 2003; Krutz and LeBeau 2006). Indeed, one of the original assumptions in the garbage can model is that a “fixed number, m, of choices is assumed. Each choice is characterized by… an entry time, the calendar time at which that choice is activated for decision” (Cohen, March, and Olsen 1972, 3). An exit time for a choice is not discussed. Research on legislative effectiveness contends that part of lawmaking often means slowly moving an idea through Congress and involves intermittent procedural victories along the way (Volden and Wiseman 2014).

This focus on ideas lingering elides a basic fact: most eventually leave Congress’s agenda not as part of a law but because they are no longer introduced. Even though disappearing in this way is the most common outcome, political scientists have spent much less effort examining failed ideas. Usually, measures that are not enacted are treated as the comparison case for successful bills. In this way, political scientists constructed a comprehensive literature that catalogs the various ways members can increase the likelihood an idea advances through Congress, the importance of the broader political environment, and measuring how effective they are at this task (Anderson, Box-Steffensmeier, and Sinclair-Chapman 2003; Craig 2021; Volden and Wiseman 2014).

A notable exception is research on messaging legislation, which identifies ideas that are introduced with the goal of making a political point (Gelman 2020; Lee 2016). Although no exact count of these bills exists, they do not describe a typical legislative proposal and are not always destined to leave the agenda. Many join the broader policy stream as parties in Congress try to enact them once political conditions change (Gelman 2017, 2020).

Although not the focus of most lawmaking and agenda-setting research, political scientists have considered reasons for why policy ideas disappear from the legislative agenda. In most cases, these are the converse for why a proposal is introduced or lingers. Kingdon (2003) speculates that measures die for a few reasons. One is that unfavorable legislative conditions, like divided government or increasing polarization, lead to a member dropping a proposal. In his interviews, he finds that sentiment reflected on Capitol Hill. One health policy staffer told him that ideas were regularly dropped because “[the staffer] felt he didn’t have the votes” in committee or on the floor (Kingdon 2003, 139).

Another reason is that policy windows close. Theories of congressional problem-solving, due to exogenous focusing events or internally imposed deadlines (Adler and Wilkerson 2012), suggest that many ideas stop being considered because Congress addressed the problem and moved on. In these situations, members introduce ideas because they know an issue is being dealt with. Once legislation is enacted, lawmakers drop the proposals that were not included. Kingdon (2003) argues this is the main reason ideas, viewed as faddish or dealing with an already solved problem, fade off the agenda. Others offer other reasons. For instance, Wawro (2004) contends that members act as policy entrepreneurs to earn promotions to leadership roles and logically, stop offering new ideas when they no longer seek them.

These various claims do not satisfactorily explain why an idea disappears if it is nearly costless to reintroduce. Even if it has little chance of enactment, why not keep a cheap measure on the agenda and accrue position-taking benefits or wait for the next policy window to open? The problem to explain is why legislators let ideas die that they have invested time and effort developing. My theory emphasizes a more complex trade-off, where members weigh keeping an old proposal on the agenda versus investing in a new one.

A Theory of Why Ideas Die

The decision to introduce a bill is often evaluated through a member’s cost-benefit analysis (Schiller 1995). They propose a measure as long as doing so is profitable. Costs include the effort it takes to write the legislation, the opportunity costs of ignoring other issues, and reputational risks back home and on Capitol Hill. Benefits include position-taking on a popular issue, receiving campaign donations from supportive interest groups, solving a policy problem, and gaining favor with important politicians like the president (Rocca and Gordon 2009; Schiller 1995; Wawro 2004). These benefits are both immediate and may accrue in the future if a measure is eventually enacted (Kingdon 2003). In other words, the expected value of introducing an idea includes immediate and prospective utility based on the likelihood the proposal becomes law.

Although the benefits for introducing legislation receive the most attention, the nature of the costs are important too. They come in two forms. The first are upfront ones associated with developing the idea. These involve the policy and political processes of taking it and turning it into legislative language. From a policy perspective, it means understanding the problem that needs to be solved, gathering relevant information, and writing the measure (Jones and Baumgartner 2005). From a political perspective, it means engaging with relevant interest groups, constituents, and fellow lawmakers to build a supportive coalition (Craig 2021; Hula 1999; Sabatier 1988). These costs can range from very small to substantial, but broadly speaking, they only need to be paid once. After the initial introduction, they are sunk.

The second type are maintenance costs that must be paid every time an idea is reintroduced. These include ensuring the measure is viewed as a viable solution to a potential problem as well as maintaining or increasing support for it in the relevant policy community. In agenda-setting parlance, these are the costs associated with softening the ground, building a favorable policy image, coalition maintenance, among others (Baumgartner and Jones 2009; Kingdon 2003).² They are paid every time an idea is (re)introduced to the agenda.³

These different costs and benefits present a more nuanced calculus for keeping an idea alive or not. Lawmakers must weigh the value of keeping the proposal on the agenda, its benefits minus the maintenance costs, and compare it to the value of introducing a new idea. If a lawmaker lets a measure die, it is more profitable to stop maintaining it and instead, develop a new one and pay its sunk costs. This willingness to walk away from a proposal varies among lawmakers. In particular, those who face lower costs or expect higher benefits when introducing something new are more likely to let their old ideas die.

On the cost side, committee leaders face lower costs in writing new legislation. Since they employ more subject matter experts, it is easier for them to produce novel material (Lewallen 2020; Schiller 1995). Moreover, given their institutionally powerful positions, interest groups provide them more legislative subsidies (Hojnacki and Kimball 1998). They also gain more expected benefits from introductions. As agenda-setters, they can increase the likelihood their idea moves through the legislative process and at a minimum, have better information about whether a measure might advance in the current political environment (Berry and Fowler 2018). For committee leaders, lower costs are limited to issues covered by their committee’s jurisdiction. These measures are cheaper to write, due to committee staff and interest group subsidies. Their agenda-setting powers increase their expected benefits. Ideas handled by other committees do not include the same advantageous cost-benefit opportunity.

Put differently, for a committee chair, it is cheap to write proposals in their jurisdiction. He has dedicated subject matter experts on staff whose effort is subsidized by organized interests. The benefits for proposing a measure are also likely higher. His idea has a better chance of enactment because he possesses agenda-setting powers and intra-party influence that help him advance it through the committee and the floor.

If his proposal fails in Congress t, he has two choices. He can persist and reintroduce the measure, even though his formidable influence did not help him secure its passage in the previous term. The cost of doing so is remarshalling or expanding the political coalition that supports his idea. Or, he can recognize that proposal’s enactment is unlikely and move on to a new one that has better prospects. Switching course is cheap, while the benefits from a new measure are likely higher.

The same advantages are not present for ideas referred to other committees. The chair lacks the same staff, interest group support, and influence over its progress. Most of its value comes from position-taking, not prospective enactment. As such, offering something new may not be more valuable and comes with steeper costs. Consequently, these institutionally advantaged lawmakers are more likely to let their ideas die between terms.

Similarly, rank-and-file members’ unenacted proposals that are referred to the committees they sit on are also more likely to leave the agenda next term. Legislators, and their staffs, have more expertise on committee-related issues and receive support from access-seeking interest groups (Curry 2019; Powell and Grimmer 2016). This lowers the cost of writing new legislation. On the benefits side, members often sit on committees because their constituents have more demand for policy in the panel’s jurisdiction (Adler and Lapinski 1997). As such, offering new measures on those topics provides substantial position-taking value. This does not explain why they would move on to something new rather than continuing to push a previously failed idea. However, their committee membership provides them a better grasp of the issues the panel might address. Their connections with committee leaders and staff provide reliable information about an idea’s prospects. This allows them to move on from moribund proposals more quickly, as they know what does or does not have a chance of providing future policy and credit-claiming benefits. In doing so, they can maximize their utility from their legislative portfolios by offering new measures with a better chance of enactment on their district’s high-demand issues.

Another reason an idea dies is its sponsor leaves Congress. Others can keep it on the agenda, but maintenance costs are very high. Even though the legislative language may not change, the new legislator needs to learn the specific idea’s politics, tend to its coalition, and set up the messaging infrastructure to advocate for it. The descriptive representation literature suggests that in some cases paying these maintenance costs may prove unrealistic. Some measures are proposed because lawmakers’ personal backgrounds help them identify an unaddressed issue and make them a particularly effective advocate (Peay 2020; Swers 2001; Volden and Wiseman 2014). The implication is that when members leave Congress, some of their ideas go with them because they are uniquely suited to pay a proposal’s maintenance costs.

Based on this argument, two main hypotheses follow. First, sections sponsored by those more advantaged in the lawmaking process—committee leaders and members whose ideas are referred to committees they lead and sit on—are more likely to die. Second, when a member leaves Congress, it increases the likelihood the proposal they sponsored leaves as well.

In addition to these hypotheses, researchers have proposed other reasons members keep or discard their policy ideas. Scholars argue that legislators create momentum for them by increasing their visibility and building support on the other side of Capitol Hill with the goal getting the measure enacted in future terms. To accomplish this, they send “Dear Colleague” letters, seek out cosponsors, and have companion bills introduced (Craig 2021; Harbridge 2015; Kirkland and Kroeger 2018). Signs of momentum increase a proposal’s expected future benefits and make it more likely that it is kept alive.

Additionally, I consider, and control for, other explanations related to the likelihood Congress will legislate in the next term. If Congress has recently worked in an area, it usually does not revisit the topic for some time (Adler and Wilkerson 2012). If a member believes a policy window has closed, an idea’s future enactment benefits disappear. Second, models of lawmaking emphasize that worsening policymaking conditions, such as divided government or a widening gridlock interval, reduce enactment opportunities (Binder 2003; Lee 2015). Consequently, ideas should be more likely to die if their sponsors do not see a path towards enactment because a less favorable legislative environment decreases the prospective benefits of it becoming law.⁴

Data and Empirical Strategy

I explore these hypotheses using data on introduced policy ideas in the US House and Senate from 1993 until 2016 (103^rd through the 114^th Congresses). Rather than focus on bills, I examine legislative sections. As recent research highlights, bills are fungible amalgamations of numerous policy ideas that are combined, broken up, and reassembled into politically palatable combinations (Casas, Denny, and Wilkerson 2020; Krutz 2001; Sinclair 2016). Usually, this restructuring occurs at the section level, meaning they can be used as a more consistent conceptualization of a policy idea (Wilkerson, Smith, and Stramp 2015).

My unit of analysis is the bill section and the main data set includes every section in an introduced House or Senate bill. I limit my scope to formally introduced measures, and not ones added later in the legislative process, to accurately track a policy idea’s original sponsor in a congressional term.⁵ Additionally, if an idea is introduced multiple times (i.e., there is a House and Senate version), I only include the earliest introduction. This way, every proposal has a single sponsor. I account for these other introductions, usually via a companion bill, as independent variables that measure an idea’s momentum in the legislative process.⁶ I exclude appropriations bills, as those policy ideas are about funding a program for that year, post office and other facility naming legislation, tariff schedule changes, and boilerplate sections that include standard language repeated across bills but does not create policy (e.g., definitions, authorization of appropriations, enactment dates).⁷ In total, this set of policy ideas includes 305,454 sections.

My interest is when an idea dies, meaning it is not enacted and is not reintroduced the next term. I track this process in two steps. First, in Congress t, I compare introduced sections to enacted ones and omit those that become law. Second, I compare the remaining sections to the ideas proposed in the next congressional term, Congress t+1. An idea dies when it is not enacted and is not reintroduced in the subsequent congress. I count an idea as being reintroduced if it arises anywhere in the legislative process in Congress t+1. Figure 1 diagrams how the legislative sections data is subset. The “reintroduced” and “not reintroduced” boxes represent the data included in my dependent variable.

Figure 1.

Flowchart describing how dependent variable is specified.

Tracking Legislative Sections Within and Between Terms

Both of these comparisons, which sections are enacted or reintroduced in the next term, are done using a text reuse approach. This sort of analysis’s underlying logic is that the same policy idea that appears in different sections shares similar language. Consequently, measures of text similarity can be used to compare how much one section resembles another. A wide-range of these statistics exist and differ in meaningful ways. Less computationally intensive measures, such as a Dice coefficient and Jaccard index, evaluate how many words two texts share but do not consider word ordering (Eatough and Preece 2021). Others, like the Smith–Waterman algorithm, consider word order (Wilkerson, Smith, and Stramp 2015). A more comprehensive approach, and the one I employ, calculates a number of similarity statistics between two legislative sections, which I detail below, and uses a machine learning model to match sections that contain the same policy idea (Casas, Denny, and Wilkerson 2020).

In this case, I trained a decision tree to recognize when two sections included the same policy idea. As a first step, two human coders hand coded 5000 section pairs to determine whether policy ideas in these texts were the same. Details on how the training set was constructed, coding rules, and inter-coder reliability statistics are included in the online appendix. The result, a dichotomous variable measuring whether two section pairs feature the same proposal, is the dependent variable used to train the decision tree. Next, I cleaned each section using standard methods. This included removing white space, headers, footers, bill titles, table of contents, punctuation, and stop words, as defined by the nltk corpus and Wilkerson, Smith, and Stramp (2015). In an effort to retain semantic differences between similar but not identical policy ideas, words were not stemmed.

Using this cleaned corpus, every section pair was compared using nine similarity statistics, which are described in Table 1. These include some that do not consider word order and others that do. They constitute the independent variables used in the decision tree. Each is calculated using n-grams, which are the number of words next to one another in a section. For instance, the title to section 105 of the PATRIOT Act is “Expansion of National Electronic Crime Task Force Initiative.” After processing the section text, the title would read “expansion national electronic crime task force initiative.” The first 2-gram is “expansion national,” and the first 5-gram is “expansion national electronic crime task.”

Table 1.

Similarity Statistics Used for Decision Tree Model.⁸

Proportion of shared 5-grams^a	The proportion of 5-grams in section A that are shared with section B
Proportion of shared 2-grams in the section heading*	The proportion of 2-grams in section A’s first 100 characters that are shared with the section B’s first 100 characters
Proportion of shared text*	Proportion of section A’s text that is shared with section B.
Longest shared block length	Length of the longest matching block of text shared between sections A and B.
Scope of shared 10-grams	The % of shared 10-grams in section A divided by the % of shared 10-grams in section B.
Shared 2-grams in the section heading.	Count of shared 2-grams in the words that are part of the first 100 characters in the section.

^aDenotes statistics that are measured separately for section A and section B.

I evaluated whether the decision tree was effective by randomly selecting 25 percent of the hand coded section pairs and using it as test data. I trained the algorithm on the other 75 percent of the hand-coded data and predicted whether sections in the test set included the same policy idea or not. The decision tree performs well, with 98 percent accuracy and precision as well as 99 percent recall of the test data.

Additionally, these similarity statistics allow me to conduct robustness checks that can alleviate concerns about the decision tree overfitting the data. The shared 5-grams, shared 2-grams in the section heading, and proportion of shared text variables all produce values between 0 and 1. Larger numbers reflect more similarity between sections and 1 means the texts, by that metric, are identical. By examining these statistics together, I categorize some matches as containing identical or near identical text. I define identical sections as those whose shared 5-grams, shared 2-grams in the section heading, and percentage of longest block length all equal 1. I define near identical matches as section pairs where those similarity measures all are greater than or equal to 0.95. As robustness checks, I estimate models based on reintroduced ideas that only include identical or near identical matches for a more conservative measure.

Measurement and Specification

Because I study ideas across two congressional terms, I use the following notation to describe the stage of an idea’s lifecycle. My initial cataloging of a measure is in congressional term t. When I assess if it is reintroduced or not, I am examining its fate in congressional term t+1. As an example, proposals introduced in the 103^rd Congress are denoted t and when I measure if they are reintroduced in the 104^th Congress, I use the t+1 notation.⁹

My main dependent variable is a dichotomous measure of whether a section that did not become law dies (1) or was reintroduced (0) in the next congressional term, t+1. Additionally, I specify two bill-level dependent variables. First, I calculate the percentage of sections from a bill that are not reintroduced in the next term. For instance, if a piece of legislation has ten sections and only three are reintroduced, this variable’s value is 0.7 to reflect that 70 percent of that legislation disappeared from the agenda. Second, I create a dichotomous variable for whether all the sections in a bill die (1) or any are reintroduced in the next congressional term (0). These variables reflect other ways to conceptualize an idea dying. If members perceive their proposals as bills and not sections, then the percentage variable captures how much of it a member decides to drop or change. The dichotomous measure of whether any bill sections survive is a more conservative test of who is more likely to let their ideas disappear.

Even though theories of agenda-setting emphasize the importance of long-term policy development in Congress, we do not know the percentage of ideas that persist on the agenda from term to term. The answer is that most are not reintroduced—only 41 percent appear in the next congressional term. An even more conservative measure that uses near identical language matches between sections lowers that value to 23 percent.

As Table 2 shows, though most sections are not reintroduced, bills they are components of remain on the agenda. Fifty percent of bills are entirely discarded. Of that 50 percent, 61 percent (or 31 percent of all bills in the dataset) only have one section. In other words, three out of every ten bills are single section measures that disappear from one term to the next. For multi-section bills, most members include some reintroduced proposals along with new ones. Fifty-eight percent have at least one section reintroduced. On average, a member drops 57 percent of the sections and retains the other 43 percent.

Table 2.

Percentage of Policy Ideas Not Reintroduced in Next Congress.

	All Bills (%)	Multi-Section Bills (%)
% of sections not reintroduced	41	41
Avg. % of sections dropped from reintroduced bill	57	57
% of bills where all sections die	50	42

Put differently, ideas can die when lawmakers stop introducing an entire bill or drop certain portions when revising a piece of legislation. Importantly, my hypotheses are the same for both processes. More advantaged legislators should be willing to pay new, sunk costs to rewrite their legislation, just as they are more willing to write entirely new bills.

Both the size of a member’s legislative portfolio, as well as the proportion they let die between terms varies a great deal. The median number of ideas proposed in a term is 111 and ranges from 1 to 1126. Figure 2 is a histogram that plots the proportion of sections that are dropped in the next term on the x-axis and the percentage of members on the y-axis. Interestingly, most portfolios mostly consist of new proposals each term. Although some are mainly reintroduced ideas, it is not the norm.

Figure 2.

Percentage of legislative portfolios not reintroduced, by member.

Moreover, the proportion of legislative portfolios that members drop between terms is fairly consistent over time. Figure 3 is a box and whiskers plot that shows both the median proportion of measures not reintroduced as well as the data’s distribution by term. Although there are fluctuations, there is no consistent increase or decrease in the median percentage of members’ portfolios that die during the 20 years I analyze.

Figure 3.

Median proportion and distribution of legislative portfolios not reintroduced, 1995–2016.

Taken together, these descriptive statistics show substantial differences in the size of lawmakers’ legislative portfolios and their willingness to stop reintroducing measures. However, in the aggregate, every term’s legislative agenda mostly consists of new ideas, not retreads.

Independent Variables

I include three variables to test my hypotheses that members with more resources to navigate the legislative process are more likely to let their ideas die. I specify dichotomous measures for whether a legislator was the chair or ranking member on a committee the measure was referred to (Chair on Referred Committee_t+1, Ranking Member on Referred Committee_t+1). I also include a dummy variable for whether the member served on a standing committee its bill was referred to (Member on Referred Committee_t+1). These variables are coded based on a member’s position and committee assignments when they decided to reintroduce an idea or not. I expect all three variables to have positive coefficients. I also include Committee Chair_t+1 and Ranking Member_t+1 dummy variables. These control for ideas proposed by chairs and ranking members that are referred to other committees. I expect each to have a negative coefficient.¹⁰

To test how closely members are tied to their ideas, I included a dichotomous measure for whether a lawmaker Returns to Congress_t+1. I hypothesize that this variable should have a negative coefficient, meaning an idea is more likely to leave the agenda if the sponsor does not return to Capitol Hill.

I measure an idea’s momentum in Congress t in three ways. First, I count the Number of Cosponsors_t that signed onto the bill. Second, I code whether the section was introduced in the other chamber in Companion_t legislation. Third, I create a dichotomous measure for whether the section Advanced_t past being referred to a committee. As committees winnow most bills, gaining attention beyond that point suggests the idea is not a pet project but has supporters pushing for its enactment. I expect all three variables to have negative coefficients.

To assess if Congress’s recent work on an issue closes a policy window and pushes ideas on the same topic off the agenda, I use the Policy Agenda Project’s data that codes each bill by major topic area and linked that topic code to sections within bills. Next, I count the Log Number of Previously Enacted Sections_t on that topic to capture Congress’s prior activity in that area. If more activity in term t is associated with closing policy windows then this variable should produce a positive coefficient. This would indicate ideas on the same topic are less likely to be reintroduced in the next term.

I account for the changing political environment by creating variables that reflect shifts between unified and divided government as well as the gridlock interval. How these political conditions might matter is about change—for instance moving from unified to divided government means fewer ideas can pass. I specify a ΔDivided Government variable, coded as −1 if government moves from unified to divided control, 0 if there is no change, and 1 if the switch is from divided to unified government. Additionally, I calculate the ΔGridlock Interval between congresses.

Finally, I include several member, section, and bill-specific control variables to account for other factors that might affect ideas being reintroduced. At the member-level, I include measures for whether they are in the Majority Party_t+1, the Number of Terms Served_t+1, and the Number of Bills Sponsored_t+1. I also include each lawmaker’s Benchmark Legislative Effectiveness Score_t, which is the lagged value from the term prior to them choosing to let a measure die (Volden and Wiseman 2014). This reflects that knowing when to keep a measure alive or let it die might be a skill or strategy that more effective legislators employ. By using this lagged benchmark variable, I ensure it is not endogenous to factors related to a legislator’s ability to introduce measures in Congress t or reintroduce them in Congress t+1.¹¹ Additionally, I include the legislator’s Ideological Extremity_t+1, which I measure with DW-NOMINATE scores and is their absolute ideological distance from the chamber median. Those outside the ideological mainstream may be more likely to keep their messaging legislation on the agenda for position-taking reasons, even though it has no chance of enactment.

At the section level, I count the number of characters in the cleaned sections to measure Section Complexity_t. Although the maintenance costs for more complicated ideas are not necessarily higher than simpler ones, the original sunk costs are. One possibility is members are more attached to the measures that took more effort to initially write. Evidence for this behavior, in which legislators fall prone to the sunk cost fallacy, would manifest if the coefficient on this variable was negative. At the bill-level, I include the Number of Sections in the Bill_t to control for potential differences in proposals that stand alone or are linked with other ideas in a broader package of policies.

I use linear probability models with member-specific random intercepts and robust standard errors. This approach allows me to conduct a between legislators analysis while acknowledging that they vary in their baseline willingness to reintroduce ideas. In contrast, a pooled model ignores this variation, a theoretically difficult position to maintain. Empirically, a pooled analysis may also understate differences across members.¹² Table 3 presents the main results. Robustness checks, including specifications that use the identical and near identical section matches, can be found in the online appendix.

Table 3.

Likelihood Policy Idea Dies in Next Congress.

	(1)	(2)	(3)
	Bill Section	% Sections Dies	Entire Bill Dies
Chair on Referred Committee_t+1	0.10**	0.11**	0.09**
Chair on Referred Committee_t+1	(0.02)	(0.02)	(0.02)
Ranking Member on Referred_t+1	0.02	0.03	0.02
Ranking Member on Referred_t+1	(0.03)	(0.02)	(0.02)
Member on Referred Committee_t+1	0.05**	0.04**	0.03**
Member on Referred Committee_t+1	(0.01)	(0.01)	(0.01)
Committee Chair_t+1	−0.01	−0.02	−0.02
Committee Chair_t+1	(0.02)	(0.01)	(0.01)
Ranking Member_t+1	0.02	0.02	0.01
Ranking Member_t+1	(0.02)	(0.01)	(0.01)
Majority Party_t+1	−0.02	−0.01	−0.01
Majority Party_t+1	(0.02)	(0.01)	(0.01)
Returning to Congress_t+1	−0.08**	−0.12**	−0.12**
Returning to Congress_t+1	(0.01)	(0.01)	(0.01)
Companion_t	−0.07**	−0.06**	−0.07**
Companion_t	(0.01)	(0.01)	(0.01)
Advanced_t	−0.09**	−0.08**	−0.14**
Advanced_t	(0.01)	(0.01)	(0.01)
# of Cosponsorst	−0.00**	−0.00**	−0.00**
# of Cosponsorst	(0.00)	(0.00)	(0.00)
Change in Divided Gov't	−0.07**	−0.04**	−0.04**
Change in Divided Gov't	(0.01)	(0.01)	(0.01)
Change in Gridlock Interval	−0.50**	−0.31**	−0.26**
Change in Gridlock Interval	(0.07)	(0.04)	(0.04)
Log Number of Previously Enacted Sections_t	−0.01	0.01**	0.01**
Log Number of Previously Enacted Sections_t	(0.01)	(0.01)	(0.01)
Terms Served_t+1	−0.01**	−0.01**	−0.01**
Terms Served_t+1	(0.01)	(0.01)	(0.01)
Number of Bills Sponsored_t+1	0.01	0.01	0.01
Number of Bills Sponsored_t+1	(0.01)	(0.00)	(0.01)
Legislative Effectiveness_t	0.01	0.01	0.01
Legislative Effectiveness_t	(0.01)	(0.01)	(0.01)
Ideological Extremity_t+1	−0.16**	−0.14**	−0.14**
Ideological Extremity_t+1	(0.05)	(0.02)	(0.02)
Senate_t+1	−0.09**	−0.03**	−0.04**
Senate_t+1	(0.02)	(0.01)	(0.01)
Log Section Complexity_t	−0.01**
Log Section Complexity_t	(0.01)
Bill Section Count_t+1	0.01	0.01**	−0.01**
Bill Section Count_t+1	(0.01)	(0.01)	(0.01)
Constant	0.87**	0.78**	0.77**
Constant	(0.04)	(0.02)	(0.02)
N	305454	75058	75058
R²	0.03	0.04	0.07
ρ	0.08	0.03	0.02

**p < 0.01; *p < 0.05.

Note: All three models are OLS regressions calculated with member-level random intercepts and robust standard errors. Model 1’s DV is whether a policy idea died or not. Model 2’s DV is the % sections of a bill that die. Model 3’s DV is whether all the sections in a bill die.

Results

Table 3 presents the results from both the section and bill-level models. Overall, the results support the theory’s main prediction—those with more lawmaking advantages are more likely to let their ideas die. A proposal introduced by a committee chair and referred to their committee is 10 percent more likely to die than ones proposed by a chairperson but sent to other committees. Similarly, a member is five percent more likely to stop reintroducing a measure if it is referred to a committee they sit on. Ranking members do not tend to let their ideas disappear more often even when it is referred to their committees.

To better conceptualize the magnitude of these differences, in Figure 4 I plot the predicted probabilities a legislator does not reintroduce a measure in the next Congress based on their committee roles. The probability a chair lets a proposal die that was previously referred to their committee is 0.73. If the legislation goes to another committee, that probability is 0.58. Similarly, the probability rank and file members do not reintroduce a measure initially referred to their committees is 0.64. When their legislation is sent to another committee, that probability drops to 0.59. I also find evidence that when members leave Congress, their ideas are more likely to leave with them. For departing legislators, the probability their measures die increases 0.08.

Figure 4.

Predicted probability a measure is not reintroduced.

In the more conservative specification, where the dependent variable is whether all the sections in a bill die, I observe similar patterns. Members returning to office are 12 percent less likely to let an entire bill disappear. However, chairs and members whose measures are sent to their committees are nine and three percent more likely to drop an entire piece of legislation. Moreover, my results are robust when I guard against false positives by subsetting on identical and near identical section matches.¹³ Finally, these findings are not arising because some members are proposing more bills, which allows them to drop portions of their expansive legislative portfolios. Not only do I control for the number of bills a member sponsors, but that variable indicates that more sponsorships is not associated with keeping an idea alive.

One concern in interpreting these results is the substantial number of factors included. The full model includes 20. To guard against interpreting potential false positives in my results, I calculate a new threshold of statistical significance using the Bonferroni correction (p < 0.0025). Even after adopting this more conservative threshold, my interpretation of significant covariates does not change.

Do Deaths Predict Introductions?

A corollary to my theory is that the number of new ideas a member introduces should be associated with how many measures they let die. This reflects part of the cost-benefit analysis I previously propose: when members abandon measures, they respond by offering new ones. Moreover, my argument suggests that those with more legislative advantages should introduce more new proposals than others.

To examine this possibility, I create a dataset where the unit is the member-congress-institutional role triad. For example, the data includes three observations for Senator Arlen Specter in the 109^th Congress. One accounts for Specter’s role as chair of the Senate Judiciary Committee. A second accounts for his role as a member on other Senate committees. The third accounts for measures that were referred to committees he did not serve on. Structuring the data in this way allows me to assess if Specter is more likely to replace failed measures in term t with new ones in term t+1 in situations where he is institutionally advantaged.¹⁴

The dependent variable is the count of how many new sections a member introduces in Congress t+1. In Specter’s case, there are three separate counts—one for how many new measures were referred to the Senate Judiciary Committee, a second for how many were referred to the other committees he sat on, and the third for those sent to committees he did not serve on. These exclude ideas that the lawmaker reintroduces and boilerplate sections. My main independent variable is the number of sections the legislator proposed in Congress t but let die in Congress t+1 (# of Dead Ideas_t+1). I expect a positive association between these two variables.

An alternative reason someone may introduce new ideas is because their measures were enacted in the previous term (Enactments_t). To account for this possibility, I count the number of sections they introduced in the previous term, again accounting for their institutional role at the time it became law. If lawmakers are replacing enactments with new introductions, I expect this coefficient to be positive.

I control for factors that may predict how many ideas a legislator introduces. These include their institutional roles as chair or ranking member on the referred committee, a member on the referred committee, serving as a committee chair or ranking member more generally, seniority, whether they are in the majority party, their legislative effectiveness, their ideological extremity, and the total number of bills they sponsor. All of these variables are operationalized in the way I describe in my previous analysis. To assess if failed ideas are associated with new legislative introductions, I specify a negative binomial regression with member and congressional term fixed effects (Table 4)

Table 4.

Likelihood Deaths are Associated With Newly Introduced Ideas.

# of Dead Ideas_t+1	0.01^**
# of Dead Ideas_t+1	(0.01)
Enactments_t	−0.01^**
Enactments_t	(0.01)
Chair on Referred Committee_t+1	0.71^**
Chair on Referred Committee_t+1	(0.05)
Ranking Member on Referred _t+1	0.27^**
Ranking Member on Referred _t+1	(0.02)
Member on Referred Committee_t+1	0.36^**
Member on Referred Committee_t+1	(0.06)
Committee Chair_t+1	−0.26^**
Committee Chair_t+1	(0.04)
Ranking Member_t+1	−0.03
Ranking Member_t+1	(0.04)
Ideological Extremity_t+1	−0.35^**
Ideological Extremity_t+1	(0.10)
Majority Party_t+1	0.07
Majority Party_t+1	(0.04)
Legislative Effectiveness_t	−0.02
Legislative Effectiveness_t	(0.01)
Senate_t+1	0.10^**
Senate_t+1	(0.04)
Terms Served_t+1	−0.04^**
Terms Served_t+1	(0.00)
Number of Bills Sponsored_t+1	0.04^**
Number of Bills Sponsored_t+1	(0.00)
Constant	0.17^*
Constant	(0.08)
N	7,269
Congress fixed effects	√
Member fixed effects	√

Note: The model is a negative binomial regression. The DV is a count of a member’s newly introduced legislative ideas.

As expected, more failed proposals from the previous congress are associated with a member introducing more new ideas in the subsequent one. The relationship is not a one-to-one exchange and varies by institutional role. Holding the number of failed ideas constant, committee leaders introduce more proposals in their areas of jurisdiction relative to less advantaged lawmakers. When he stops reintroducing an average number of sections, 21, a committee chair will introduce 1.2 new measures that are referred to his committee. A ranking member will propose 0.8, a member of the committee will sponsor 0.5, and someone not on the committee will offer 0.2. Surprisingly, enacting more sections in the previous term does not spur new introductions. The small substantive effects from this model suggest other factors affect the size and scope of members’ legislative portfolios. Yet, even after controlling for a number of factors, when they stop introducing proposals, legislators offer new ones.

Conclusion

The most common fate for a policy idea on Capitol Hill is not enactment but a quiet death where it is no longer introduced. The circumstances in which a proposal stops being viewed as a viable legislative alternative are not well known but important for understanding the policymaking process as well as how members advocate for their ideas in Congress.

I argue that when deciding to drop a measure from their portfolio, members consider whether it is worth proposing a new idea instead. This means weighing their benefits and different costs. Reintroduced proposals come with maintenance costs while new ones require investment in both idea development and building a supportive coalition. Consequently, the legislators that are most likely to let an idea die are those who can most afford proposing a fresh measure and have better information about a proposal’s future enactment prospects. In other words, these are legislators for whom the costs of policy development and the benefits of a lingering proposal are lower.

My account of why members let ideas disappear gives rise to several hypotheses that I test using data on proposals reintroduced or not from 1993 through 2016. I find that lawmakers are more likely to let the policy ideas they have expertise in and jurisdiction over die between congressional terms. A departing member, as expected, reduces the likelihood a measure remains on the agenda. These findings hold after considering a number of alternatives, including the confounds of members who introduce more bills in the first place, a policy window closing, and changing lawmaking conditions.

This analysis raises questions about how the policymaking process is understood. If most policy ideas are proposed, not enacted and die, how are we to best understand the bulk of Congress’s agenda? The long-term process described by many policy process theories does not apply to most measures. Then, are most proposals one-shot attempts at lawmaking, messaging for interest groups and attentive publics, or a more scattershot process where members propose ideas and see what sticks (i.e., a spaghetti against the wall model instead of a garbage can model)? More analysis of this typical lifecycle, instead of the exceptional, will help answer some of these questions.

Moreover, political scientists have studied, and often bemoaned, weakened committee power, the cuts to policymaking staff, and the end of effective subcommittees. Even in an era of stronger parties, my analysis points to an important way that committee membership affects how lawmakers work. Knowing when to quit and refocus one’s attention is affected by resources and insider knowledge. Empirically, this increases the legislative agenda’s churn. Normatively, it may improve Congress’s responsiveness. Although this data cannot speak directly to that question, scholars should explore this possibility. If representation is enhanced by agenda turnover, then an even more compelling argument can be made that reinvigorating subcommittees or hiring more policymaking staff will improve the institution.

Furthermore, this analysis points to future research that can further explain why some policy ideas are durable and others are not. Theoretically, this paper points to an understudied latent concept of policy viability on Capitol Hill. Why legislators stop seeing most measures as workable is a consequential question that this paper highlights but is not equipped to answer. Explaining this phenomenon will move forward the policymaking and agenda-setting literatures. Additionally, a key part of my argument is that more staff resources make it easier for members to drop ideas and move onto new ones. I also show that measures sometimes persist on the agenda once a legislator leaves office. As staff do much of the legwork in producing and advocating for policies, it is possible they take their measures with them as they move offices. Connecting staff with original introductions and tracking this movement can provide important insight into the role staff play as policy entrepreneurs.

This paper is limited in that I only track ideas reintroduced between two congressional terms. I acknowledge this limitation, but given the remarkable churn of Congress’s agenda, it is likely only a very small subset of ideas survives over the long term. Even so, studying this group of measures, and what makes them exceptional, will be a valuable effort.

From a scope and methods standpoint, I lean into a big data framework. Both the data and my connecting sections between congresses can be adapted to answer other important questions like who is responsible for introducing new ideas, estimating why some ideas linger on the agenda for a long time, or studying influence on specific issues. Different approaches will also be valuable in extending many of the ideas outlined in this paper. Interviews and participant observation can provide new insights into how policymakers understand policy viability. More targeted analyses at specific issue areas or legislative moments might explain why so many ideas disappear from the agenda, why lawmakers are not responding to closing policy windows, and the importance of momentum. By combining methods that allow researchers to more closely track an idea’s progression through Congress with other approaches, political scientists can begin to address numerous outstanding questions on lawmaking and agenda-setting.

Supplemental Material

Supplemental Material - The Deaths of Ideas in Congress

Supplemental Material for The Deaths of Ideas in Congress by Jeremy Gelman in Political Research Quarterly

Footnotes

Acknowledgements

I thank Larry Evans, Leah Rosenstiel, Josh Ryan, Christopher Wlezien, and participants at the 2022 American Political Science Association conference, the 2023 Midwest Political Science Association conference, the UNR Political Science Faculty Research workshop, and four anonymous reviewers for their helpful comments. I also thank Karen Simpson and Reilly Hogan for their research assistance.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Center for Legislative Effectiveness.

ORCID iD

Jeremy Gelman

Data Availability Statement

Replication files for the paper are available at .

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Adler

E. Scott

Lapinski

John

. 1997. “Demand-Side Theory and Congressional Committee Composition: A Constituency Characteristics Approach.” American Journal of Political Science 41 (3): 895–918.

Adler

E. Scott

Wilkerson

John

. 2024. Congressional Bills Project: 1993-2016. https://www.congressionalbills.org/

Adler

E. Scott

Wilkerson

John

. 2012. Congress and the Politics of Problem Solving. New York: Cambridge University Press.

Anderson

William

Box-Steffensmeier

Janet

Sinclair-Chapman

Valeria

. 2003. “The Keys to Legislative Success in the U.S. House of Representatives.” Legislative Studies Quarterly 28 (3): 357–386.

Baumgartner

Frank R.

Jones

Bryan D.

. 2009. Agendas and Instability in American Politics. 2nd ed. Chicago, IL: University of Chicago Press.

Berry

Christopher

Fowler

Anthony

. 2018. “Congressional Committees, Legislative Influence, and the Hegemony of Chairs.” Journal of Public Economics 158: 1–11.

Binder

Sarah

. 2003. Stalemate: Causes and Consequences of Legislative Gridlock. Washington DC: Brookings Institution Press.

Casas

Andreu

Denny

Matthew J.

Wilkerson

John

. 2020. “More Effective Than We Thought: Accounting for Legislative Hitchhikers Reveals a More Inclusive and Productive Lawmaking Process.” American Journal of Political Science 64 (1): 5–18.

Cohen

Michael D.

March

James G.

Olsen

John P.

. 1972. “A Garbage Can Model of Organizational Choice.” Administrative Science Quarterly 17 (1): 1–25.

10.

Craig

Alison W.

2021. “It Takes a Coalition: The Community Impacts of Collaboration.” Legislative Studies Quarterly 46 (1): 11–48. DOI: 10.1111/LSQ.12324

11.

Curry

James

. 2019. “Knowledge, Expertise, and Committee Power in the Contemporary Congress.” Legislative Studies Quarterly 44 (2): 203–237.

12.

Eatough

Mandi

Preece

Jessica

. 2021. “Toward a Fuller Account of Lawmaking: The Lawmaking Productivity Metric (LawProM).” Working Paper. https://www.womeninlegislativestudies.org/uploads/1/3/5/1/135128342/eatough_and_preece_wils_research_seminar.pdf

13.

Gelman

Jeremy

. 2017. “Rewarding Dysfunction: Interest Groups and Intended Legislative Failure.” Legislative Studies Quarterly 42 (4): 661–692.

14.

Gelman

Jeremy

. 2020. Losing to Win: Why Congressional Majorities Play Politics Instead of Make Laws. Ann Arbor, MI: University of Michigan Press.

15.

Griffin

John D.

2014. “When and Why Minority Legislators Matter.” Annual Review of Political Science 17: 327–336.

16.

Harbridge

Laurel

. 2015. Is Bipartisanship Dead? Policy Agreement and Agenda-Setting in the House of Representatives. New York: Cambridge University Press.

17.

Hojnacki

Marie

Kimball

David

. 1998. “Organized Interests and the Decision of Whom to Lobby in Congress.” American Political Science Review 92 (4): 775–790.

18.

Hula

Kevin

. 1999. Lobbying Together: Interest Group Coalitions in Legislative Politics. Washington, DC: Georgetown University Press.

19.

Jones

Bryan D.

Baumgartner

Frank R.

. 2005. The Politics of Attention. Chicago, Illinois: University of Chicago Press.

20.

Kingdon

John

. 2003. Agendas, Alternatives, and Public Policies. 2nd ed. New York: Longman.

21.

Kirkland

Justin

Kroeger

Mary

. 2018. “Companion Bills and Cross-Chamber Collaboration in the U.S. Congress.” American Politics Research 46 (4): 629–670.

22.

Krutz

Glen

. 2001. Hitching a Ride: Omnibus Legislating in the US Congress. Columbus, OH: Ohio State University Press.

23.

Krutz

Glen

LeBeau

Justin

. 2006. “Recurring Bills and the Legislative Process in the US Congress.” Journal of Legislative Studies 12 (1): 98–109.

24.

Kuhlmann

Johanna

van der Heijden

Jeroen

. 2018. “What Is Known about Punctuated Equilibrium Theory? And What Does That Tell Us about the Construction, Validation, and Replication of Knowledge in the Policy Sciences?” Review of Policy Research 35 (2): 326–347.

25.

Lee

Frances

. 2015. “How Party Polarization Affects Governance.” Annual Review of Political Science 18: 326–347.

26.

Lee

Frances

. 2016. Insecure Majorities: Congress and the Perpetual Campaign. Chicago, Illinois: University of Chicago Press.

27.

Lewallen

Jonathan

. 2020. “Booster Seats: New Committee Chairs and Legislative Effectiveness.” Journal of Legislative Studies 26 (4): 496–522.

28.

Peay

Periloux

. 2020. “Incorporation Is Not Enough: The Agenda Influence of Black Lawmakers in Congressional Committees.” Journal of Race, Ethnicity, and Politics 6 (2): 402–438.

29.

Powell

Eleanor Neff

Grimmer

Justin

. 2016. “Money in Exile: Campaign Contributions and Committee Access.” The Journal of Politics 78 (4): 974–988. DOI: 10.1086/686615.

30.

Rawat

Pragati

Morris

John C.

. 2016. “Kingdon’s ‘Streams’ Model at Thirty: Still Relevant in the 21st Century?” Politics & Policy 44 (4): 608–638.

31.

Rocca

Michael S.

Gordon

Stacy B.

. 2009. “The Position-Taking Value of Bill Sponsorship in Congress.” Political Research Quarterly. 63 (2): 387–397.

32.

Sabatier

Paul

. 1988. “An Advocacy Coalition Framework of Policy Change and the Role of Policy-Oriented Learning Therein.” Policy Sciences 21: 129–168.

33.

Schiller

Wendy

. 1995. “Senators as Political Entrepreneurs: Using Bill Sponsorship to Shape Legislative Agendas.” American Journal of Political Science 39 (1): 186–203.

34.

Sinclair

Barbara

. 2016. Unorthodox Lawmaking: New Legislative Processes in the U.S. Congress. Washington, DC: Congressional Quarterly Press.

35.

Stewart

Charles.

2021a. “House Standing Committee Assignments, 103rd - 117th Congress.” Harvard Dataverse. DOI: 10.7910/DVN/XLIHUC.

36.

Stewart

Charles

. 2021b. “Senate Standing Committee Assignments, 103rd - 117th Congress.” Harvard Dataverse. DOI: 10.7910/DVN/EQ6KC.

37.

Swers

Michele

. 2001. “Understanding the Policy Impact of Electing Women: Evidence from Research on Congress and State Legislatures.” PS: Political Science and Politics 34 (2): 217–220.

38.

Volden

Craig

Wiseman

Alan E.

. 2014. Legislative Effectiveness in the United States Congress. New York: Cambridge University Press.

39.

Wawro

Gregory

. 2004. Legislative Entrepreneurship in the U.S. House of Representatives. Ann Arbor, MI: University of Michigan Press.

40.

Wilkerson

John

Smith

David

Stramp

Nicholas

. 2015. “Tracing the Flow of Policy Ideas in Legislatures: A TextReuse Approach.” American Journal of Political Science 59 (4): 943–956.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.37 MB